Search papers, labs, and topics across Lattice.
This paper introduces Graph Wavelet Compressed Sensing (GWCS), a novel framework that efficiently compresses graph signals by leveraging sparse wavelet-domain representations through the spectral graph wavelet transform. By integrating a nonparametric multilevel importance sampler with a scale-aware graph neural network, the method retains critical wavelet coefficients while enabling accurate signal reconstruction from sparse data. Evaluations on synthetic and PDE simulation datasets reveal that GWCS significantly outperforms existing graph signal sampling methods and autoencoders in terms of reconstruction fidelity and data compression efficiency.
Achieving high reconstruction fidelity with substantial data compression, GWCS redefines the efficiency of graph signal processing in scientific machine learning.
Scientific machine learning methods such as neural operators and physics-informed neural networks have advanced engineering applications and inverse problems, but their training typically requires large volumes of simulated data. This makes data preparation and model training expensive. We propose Graph Wavelet Compressed Sensing (GWCS), a learning-based framework for offline compression of graph signals by representing them as sparse, interpretable wavelet-domain representations using the spectral graph wavelet transform. The framework combines a nonparametric multilevel importance sampler, which retains high-energy wavelet coefficients within each scale for a given compression ratio, with a scale-aware graph neural network that reconstructs the signal from the sparse coefficients. We evaluate the proposed framework on synthetic approximately band-limited graph signals over random graphs and four PDE simulation datasets over meshes, which include Turbulent Radiative Layer, Viscoelastic Instability, Kolmogorov Flow, and Dynamic Stall. We compare against graph signal sampling methods and graph autoencoder baselines. Results demonstrate that the framework achieves high reconstruction fidelity and substantial data compression compared to existing benchmarks.
